differential expression of e-cadherin and type iv...

Advances in Brief

Differential Expression of E-cadherin and Type IV CollagenaseGenes Predicts Outcome in Patients with Stage I Non-SmallCell Lung Carcinoma1

Roy S. Herbst, Seiji Yano, Hiroki Kuniyasu,Fadlo R. Khuri, Corazon D. Bucana, Fang Guo,Diane Liu, Bonnie Kemp, J. Jack Lee,Waun Ki Hong, and Isaiah J. Fidler2

Departments of Thoracic/Head and Neck Medical Oncology [R. S. H.,F. K., F. G., D. L., B. K., W. K. H.], Cancer Biology [S. Y., H. K.,C. D. B., I. J. F.], and Biostatistics [J. J. L.], The University of TexasM. D. Anderson Cancer Center, Houston, Texas 77030

AbstractBecause routine histopathological examination of pri-

mary non-small cell lung cancer does not predict diseaseoutcome, we correlated disease outcome with the expressionlevel of multiple genes that regulate distinct steps of themetastatic process in 60 formalin-fixed, paraffin-embedded,archival specimens of stage I lung carcinoma from patientsundergoing curative surgery at the M. D. Anderson CancerCenter. The expression of E-cadherin (related to cell co-hesion), type IV collagenase [matrix metalloproteinase(MMP)-2 and MMP-9, related to invasion], and three an-giogenic molecules, basic fibroblast growth factor, vascularendothelial growth factor/vascular permeability factor, andinterleukin 8, were examined by a colorimetric in situmRNA hybridization technique. The expression levels of theindividual genes analyzed by a Cox univariate analysis werenot prognostic. In contrast, the ratio between expression oftype IV collagenases (mean of the expression of MMP-2 andMMP-9) and E-cadherin, the MMP:E-cadherin ratio (meas-ured at the periphery of each tumor), was significantlyhigher in patients with recurrent disease than in patientswho remained disease free (P5 0.00003). Longer overallsurvival and reduced disease recurrence rates were signifi-cantly associated with a lower MMP:E-cadherin ratio (<2)by a Kaplan-Meier survival analysis (P 5 0.0002 andP 50.0001, respectively). Multiple covariate analyses of overall

and disease-free survival also concluded that the MMP:E-cadherin ratio was a significant prognostic factor whencorrected for age (P5 0.0001). Determination of this geneexpression ratio in individual human lung cancers mighttherefore be used to direct tailored treatment for individualpatients with resectable lung cancer.

IntroductionLung cancer accounts for 175,000 deaths annually in the

United States, and most attributable to NSCLC3 are caused bymetastasis. The prognosis for lung cancer is best in patients withstage I disease; however, even in these patients,.40% willrelapse subsequent to surgical resection (1–4). Because detailedhistopathological examination of primary lesions cannot be usedto accurately predict disease outcome, there exists a great needto identify molecular markers with which to distinguish patientswith resectable lung cancer at risk of recurrent disease.

Several molecular prognostic factors for human lung can-cers have been proposed, including the presence of K-ras (5),the loss of blood group antigen A (6), and elevated bcl-2expression (7). None of these factors, however, have beenaccepted for routine clinical use. Two univariate and multivari-ate analyses of multiple prognostic indicators identified patho-logical stage, histological subtype, and tumor invasiveness (intolymph nodes and blood vessels) as critical prognostic factors (8,9). The presence of K-ras, the absence of p21-ras, and low levelsof bcl-2 protein, shown to be important negative factors, alsocorrelated with disease outcome (9). In addition, a recent mul-tivariate analysis of 260 patients with surgically resected stageI/II lung cancer demonstrated shorter survival for patients withoverexpression of bcl-2 antigen and Ki67 (10). In this study,tumor microvessel density did not show statistical significance(10). In other studies, however, the extent of vascular supply tolung cancers as measured by microvessel density has beenshown to correlate directly with disease stage and inversely withsurvival (11–14).

Because most patients with resectable lung cancer succumbto metastatic disease (1–4), prognostic factors based on themetastatic potential of these neoplasms should predict diseaseoutcome. The process of tumor metastasis is highly selectiveand consists of multiple, sequential events that include growth,induction of angiogenesis, detachment, invasion, adhesion, andproliferation (15). This is followed by the induction of angio-genesis at distant sites (15–17). Because each of the discrete

Received 10/28/99; revised 12/28/99; accepted 1/11/00.The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisementin accordance with 18 U.S.C. Section 1734 solely toindicate this fact.1 This work was supported in part by Cancer Center Support CoreGrants CA16672 and R35-CA42107 from the National Cancer Institute,NIH (to I. J. F.), and grants from the M. D. Anderson Physician ReferralService, the University of Texas Specialized Program of ResearchExcellence in Lung Cancer Development Award, and an AmericanSociety of Clinical Oncology Career Development Award (to R. S. H.).2 To whom requests for reprints should be addressed, at Department ofCancer Biology, Box 173, The University of Texas M. D. AndersonCancer Center; 1515 Holcombe Boulevard, Houston, TX 77030. Phone:(713) 792-8577; Fax: (713) 792-8747; E-mail: [email protected].

3 The abbreviations used are: NSCLC, non-small cell lung cancer; ISH,in situ hybridization; MMP, matrix metalloproteinase; bFGF, basicfibroblast growth factor; VEGF/VPF, vascular endothelial growth fac-tor/vascular permeability factor; IL, interleukin; IHC, immunohisto-chemistry.

790 Vol. 6, 790–797, March 2000 Clinical Cancer Research

Research. on January 10, 2020. © 2000 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

http://clincancerres.aacrjournals.org/

steps of metastasis is regulated by independent genes, the iden-tification of cells with metastatic potential in heterogeneousneoplasms requires multiparametric, multivariate analysis ofrelevant gene expression (15, 18, 19). Our laboratory has de-veloped a rapid, colorimetric, ISH technique to detect the ex-pression of specific genes that regulate the different steps ofmetastasis (including angiogenesis; Refs. 18, 20, and 21). ThisISH technique uses oligonucleotide probes to detect specificmRNA transcripts in frozen and formalin-fixed, paraffin-embedded specimens and can determine the expression level ofmultiple genes that regulate different steps of metastasis such asE-cadherin (which is related to cell cohesion/detachment), col-lagenase type IV (MMP-2 and MMP-9, which is related toinvasion), and bFGF, VEGF/VPF, and IL-8 (which are related toangiogenesis). Previous reports from our laboratory demonstratethat the analysis using these genes (but not others) predictsmetastatic potential of individual patients’ colorectal carcinoma(21–23), gastric carcinoma (24), prostate carcinoma (25), andpancreatic carcinoma (26). On the basis of these findings, wehypothesized that this panel of genes could be used effectivelyto analyze NSCLC. The current report concerns 30 cases ofstage I lung adenocarcinoma and 30 cases of stage I lungsquamous cell carcinoma from the archives of the M. D. Ander-son Cancer Center. Our analysis of the expression of metastasis-related genes and the ratio between the relative expression levelslead us to conclude that this assay can indeed be used to assessprognosis and ultimately direct specific treatment for this deadlydisease.

Materials and MethodsNSCLC Patients. Archival specimens of 33 adenocarci-

noma cases and 30 squamous carcinoma cases were identifiedby a pathologist without regard to clinical outcome data. Uponexamination of the histopathology and clinical data, three ade-nocarcinoma cases were found to represent stage IIIA disease(positive mediastinal lymph nodes). These three cases wereeliminated from the final analysis, which dealt with 30 adeno-carcinomas and 30 squamous cell carcinomas. The patient char-acteristics are listed in Table 1. Patient survival was verified and

updated through the Tumor Registry as of July 1, 1998. By thelast follow-up, 33 patients had died and 27 patients were alive.

Oligonucleotide Probes. Specific antisense oligonucleo-tide DNA probes for seven different metastasis-related geneswere designed complementary to the mRNA transcripts, basedon published reports of the cDNA sequences (18, 27–35). Thespecificity of the oligonucleotide sequences was initially deter-mined by a GenBank/European Molecular Biology Laboratorydatabase search, using the Genetics Computer Group sequenceanalysis program (Genetics Computer Group, Madison, WI) andbased on the FastA algorithm that showed 100% homology withthe target gene and minimal homology with nonspecific mam-malian gene sequences (36). The specificity of each sequencewas also confirmed by Northern blot analysis (26). A polyd(T)20 oligonucleotide was used to verify the integrity of mRNAin each sample (21–23). All DNA probes were synthesized withsix biotin molecules (hyperbiotinylated) at the 39end via directcoupling using standard phosphormidine chemistry (ResearchGenetics, Huntsville, AL; Refs. 35 and 37). The lyophilizedprobes were reconstituted to a 1-mg/ml stock solution in 10 mMTris-HCl (pH 7.6) and 1 mM EDTA. The stock solution wasdiluted with probe diluent (Research Genetics) immediatelybefore use. The working dilutions of each probe are shown inTable 2.

ISH. Tissues stained with H&E were first reviewed forthe presence of tumor cells. ISH was performed as describedpreviously using the Microprobe manual staining system (FisherScientific, Pittsburgh, PA; Ref. 26). Tissue sections (4mmthick) of formalin-fixed, paraffin-embedded specimens weremounted on silane-coated ProbeOn slides (Fisher Scientific).The slides were placed in a Microprobe slide holder, dewaxed,and dehydrated with Autodewaxer and Autoalcohol (ResearchGenetics), followed by enzymatic digestion with pepsin (35).Hybridization of the probe was carried out for 60 min at 45°C,and the samples were then washed three times with 23 SSC for2 min at 45°C (13SSC5 0.15M NaCl, 0.015M sodium citrate).The samples were incubated for 30 min in alkaline phosphatase-labeled avidin at 45°C, briefly rinsed in 50 mM Tris buffer (pH7.6), rinsed for 1 min with alkaline phosphatase enhancer(Biomeda Corp., Foster City, CA), and incubated for 30 minwith the chromogen substrate FastRed (Research Genetics) at45°C. A positive reaction in this assay stained red. Control forendogenous alkaline phosphatase included treatment of the sam-ples in the absence of the biotinylated probe and use of chro-mogen in the absence of any oligonucleotide probes. To checkthe specificity of the hybridization signal, the following controlswere used: (a) RNase pretreatment of tissue sections; (b) abiotin-labeled sense probe; and (c) a competition assay withunlabeled antisense probe. A markedly decreased or absentsignal was obtained with these treatments.

Image Analysis to Quantify Intensity of Color Reaction.Stained sections were examined in a Zeiss photomicroscope(Carl Zeiss, Inc., Thornwood, NY) equipped with a three-chipcharge-coupled device color camera (model DXC-960 MD;Sony Corp., Tokyo, Japan). The images were analyzed usingOptimas image analysis software (version 5.2; Bothell, WA).The slides were prescreened by one of the investigators todetermine the range in staining intensity of the slides to beanalyzed. Images covering the range of staining intensities were

Table 1 Patient characteristics

Characteristic No. of patients

SexMale 42Female 18

HistologyAdenocarcinoma 30Squamous cell carcinoma 30

RaceWhite 57Black/Hispanic 3

StageT1 ,3.0 cm in diameter 27T2 .3.0 cm in diameter 33

AgeRange 39–82Mean (SD) 63.7 (9.7)Median 65

791Clinical Cancer Research



captured electronically, a color bar (montage) was created, anda threshold value was set in the red, green, and blue modes ofthe color camera. All subsequent images were quantified basedon this threshold. The integrated absorbance of the selectedfields was determined based on its equivalence to the mean loginverse gray scale value multiplied by the area of the field. Thesamples were not counterstained; therefore, the absorbance wasattributable solely to the product of the ISH reaction. For eachsection, we determined the absorbance in several 23 2-mm

zones located at the periphery of the tumors or at the foci ofstromal invasion. Three to five different fields in each 232-mm zone were quantified to derive an average value by areader who was kept blinded from the clinical outcome. Theintensity of staining was standardized to that of the integratedabsorbance of poly d(T)20 and determined by comparison withthe integrated absorbance of nonpathological lung epithelium(tumor-free tissue), which was set at 100.

MMP:E-cadherin Ratio. The MMP:E-cadherin ratiowas calculated as the ratio of the expression of MMP (anaverage value of MMP-21 MMP-9) divided by the expressionlevel of E-cadherin. This ratio has been shown previously tocorrelate with prognosis in prostate (25), pancreas (26), colon(21–23), and gastric cancers (24).

Statistical Analysis. The minimum level of gene expres-sion among the groups was compared by the Wilcoxon rank sumtest. Survival probability was estimated by the Kaplan-Meiermethod (38). The prognostic effect of putative covariates ondisease-free survival and overall survival was examined by theunivariate and stepwise multiple-covariate Cox models.

ResultsExpression of Metastasis-related Genes in Stage I

NSCLC Specimens. Prior to analysis, the integrity of mRNAin each sample was verified by using a poly d(T)20 probe (22,39). All 60 samples had an intense histochemical reaction,indicating that the mRNA was well preserved. After ISH withthe different probes, a quantitative value for each probe wasdetermined by comparing the expression at the most intensely

Table 2 Sequence of oligonucleotide probes

Probe Sequence 59–39GC

content (%) Dilution Ref.

MMP-2 GGC CAC ATC TGG GTT GCG GC 70.0 1:200 (25)MMP-9 CCG GTC CAC CTC GCT GGC GCT CCG G 80.0 1:200 (27)bFGF CGG GAA GGC GCC GCT GCC GCC 87.5 1:200 (55)VEGF TGG TGA TGT TGG ACT CCT CAG TGG GC 57.7 1:200 (31, 54, 56)IL-8 CTC CAC AAC CCT CTG CAC CC 65.0 1:200 (28, 57, 58)E-cadherin TGG AGC GGG CTG GAG TCT GAA CTG 62.5 1:200 (33)(Mixture) GAC GCC GGC GGC CCC TTC ACA GTC 75.0 1:200Poly d(T)20 TTT TTT TTT TTT TTT TTT TT 1:1000 (39)

Table 3 Expression of metastasis-related genes in stage I human NSCLC

Clinicopathological variables E-cadherin MMP-2 MMP-9 Ratio n IL-8 VEGF bFGF

All cases 76.06 34.1 124.06 50.2 127.06 52.6 1.886 0.93 48 151.16 101.1 163.06 101.3 167.46 102.2Stage

T1 (n 5 27) 82.06 42.2 129.66 52.4 141.06 58.5 1.946 0.74 22 181.36 115.0a 177.06 116.0 178.46 95.0T2 (n 5 33) 71.16 25.4 119.36 48.5 115.66 45.1 1.836 1.07 26 124.46 79.4 151.26 87.5 158.16 108.9

PathologyAd (n 5 30) 69.86 35.5a 128.96 60.0 134.96 57.4 2.206 1.1a 24 182.16 129.9 174.46 123.1 164.36 101.6Sq (n5 30) 82.26 32.0 119.06 38.2 119.26 47.1 1.556 0.38 24 118.86 39.8 115.16 74.4 170.56 104.9

SexM (n 5 42) 73.56 29.0 125.76 51.6 126.56 50.0 1.956 0.71 34 132.76 73.2 149.96 82.9 178.06 113.5F (n 5 18) 81.96 44.3 120.16 47.9 128.36 59.9 1.846 1.02 14 192.76 140.4 194.96 134.4 141.66 63.9

Age,60 (n 5 19) 65.66 16.7 130.66 50.6 126.66 63.3 1.976 0.71 14 191.96 134.9 191.06 105.5 218.36 137.8$60 (n 5 41) 80.86 38.9 120.96 50.3 127.26 47.8 1.846 1.02 34 133.16 77.7 151.56 98.7 46.46 76.5

a P , 0.05 by Wilcoxon rank sum.

Table 4 Cox model: univariate and multiple-covariate analyses

Univariate analysisOverallsurvival

Disease-freesurvival

Age 0.28 0.91Stage (T1 vs.T2) 0.94 0.94Sex 0.39 0.49Histology 0.36 1.00Ratioa 0.002 0.00003E-cadherin 0.15 0.19MMP-2 0.89 0.27MMP-9 0.45 0.31IL-8 0.89 0.32bFGF 0.29 0.34VEGF 0.95 0.60

Multiple-covariate analysis

Age 0.14 0.15Ratioa 0.0001 0.0001a Ratio of [(MMP-2 1 MMP-9) 4 2] 4 E-cadherin.

792 E-cadherin and Type IV Collagenase in Lung Cancer



stained tumor area with that of normal, tumor-free lung tissue.In normal bronchial tissue, cytoplasmic staining of the metas-tasis-related genes was observed in bronchial epithelial cells,bronchial glands, type I and IV alveolar cells, and macrophages.The quantitative data are summarized in Table 3. The expressionof all of the genes under study except E-cadherin was reproduc-ibly higher in the tumor tissue than in tumor-free tissue. Nosignificant differences in gene expression were found betweenthe adenocarcinoma and squamous cell carcinoma groups. Ingeneral, the expression of the genes analyzed as a single factor

did not correlate with overall survival or disease-free survival(Table 4).

The Ratio Between Expression of MMP-2, MMP-9,and E-cadherin Predicts Disease Recurrence.The ex-pression of MMP-9 and MMP-2 was compared with theexpression of E-cadherin. The calculated ratio between thesevalues differed between the adenocarcinomas (2.206 1.1)and the squamous carcinomas (1.556 0.38). Two represent-ative cases for stage I adenocarcinomas are shown in Fig. 1,and two representative cases for stage I squamous cell car-

Fig. 1 ISH analysis for expression of E-cadherin, MMP-2, and MMP-9 in human lung adenocarcinomas and in tumor-free tissue. H&E stainingshows the tumors to be moderately differentiated adenocarcinomas. Hybridization with a hyperbiotinylated poly d(T)20 probe confirmed the integrityof mRNA (red reaction). The expression intensity for E-cadherin and MMP-2 and MMP-9 in the tumor-free tissue was assigned a value of 100, andthe measurements in the tumor tissue are relative to that number. The MMP:E-cadherin ratio defined as [(MMP-21 MMP-9) 4 2] 4 E-cadherin was1.0 (A) and 3.9 (B).Bar, 50 mm.




cinomas are shown in Fig. 2. Staining with H&E and probingwith the poly d(T) oligonucleotide provide evidence forhistological type and RNA integrity, respectively. Fig. 1Aexhibits a case of an adenocarcinoma with a low (1.0) MMP:E-cadherin ratio, and Fig. 1B shows an adenocarcinoma witha high (3.9) MMP:E-cadherin ratio. Fig. 2A shows a case ofa stage I squamous cell carcinoma with a low (1.2) MMP:E-cadherin ratio, whereas Fig. 2B exhibits a case of stage Isquamous cell carcinoma with a high (2.9) MMP:E-cadherinratio. As a whole, tumor tissue had reduced E-cadherinexpression (766 34) and increased MMP expression

(MMP-2, 124 6 50.2; MMP-9, 1276 52.6) as comparedwith adjacent normal tissue (Table 3). No significant differ-ences were observed in the mean ratios when compared forsubstage (T1 versusT2), sex, and age.

A higher MMP:E-cadherin ratio was significantly associ-ated with shorter overall survival (P5 0.002) and shorterdisease-free survival (P5 0.00003; Table 4). Furthermore,using a stepwise multiple-covariate Cox model analysis, boththe MMP:E-cadherin ratio and age were chosen. After adjust-ment for age, the MMP:E-cadherin ratio was a highly significantpredictor of decreased overall survival (P5 0.0001). Similar

Fig. 2 ISH analysis of E-cadherin, MMP-2, and MMP-9 mRNA in human lung squamous cell carcinomas and in tumor-free tissue. H&E stainingconfirms the tumor to be a squamous cell carcinoma. Hybridization with hyperbiotinylated poly d(T)20 probe confirmed the integrity of mRNA (redreaction). The expression intensity for E-cadherin and MMP-2 and MMP-9 in tumor-free tissue was assigned a value of 100. The ratio of [(MMP-21 MMP-9) 4 2] 4 E-cadherin was 1.3 (A) and 2.9 (B).Bar, 50 mm.




results were found for disease-free survival, where the MMP:E-cadherin ratio is the only significant prognostic factor.

Kaplan-Meier survival curves are shown in Fig. 3. Using acutoff value for the MMP:E-cadherin ratio of 2.0 (correspondingto the third quartile of the ratio), there was a significant differ-ence in overall survival (median, 10.2versus2.3 years;P 50.0002) and disease-free survival (median, not reachedversus1.2 years;P 5 0.0001), favoring the group with a ratio,2.0(Fig. 3,A andB, respectively). This also held true when the 30adenocarcinoma cases were evaluated independently; however,the squamous carcinoma cell subgroup analyzed alone did notachieve statistical significance.

DiscussionWe determined whether the expression level of metastasis-

related genes in stage I human lung cancer would have prog-nostic significance. Age, substage (T1 and T2), histology, andsex, as well as levels of expression of bFGF, VEGF/VPF, andIL-8, did not predict overall patient survival or disease recur-rence. However, the ratio between the expression of MMP-2,MMP-9, and E-cadherin significantly correlated with disease-free and overall patient survival.

We have reported previously that the metastatic potentialof human colon (21–23), gastric (24, 40), prostate (25), and

pancreatic (26) cancers can be identified by a multiparametricanalysis for the expression of genes that encode for epidermalgrowth factor receptor (growth), bFGF, IL-8 (angiogenesis),E-cadherin (cell cohesion), and MMP-2 and MMP-9 (invasion).Like other neoplasms, human lung cancers consist of multiplecell types, including tumor cells, fibroblasts, normal epithelialcells, endothelial cells, and infiltrating leukocytes (15). Becausemetastasis can be produced by a small subpopulation of tumorcells (,1.0% of the tumor), detecting the expression of metas-tasis-related genes requires a sensitive technique (15, 41). Wechose ISH because it identifies the cellular source of the mRNAas well as intratumoral heterogeneity in expression, whereasNorthern blot analysis represents only the average levels ofmRNA of all of the cells in a sample (21, 22).

As was the case for other human carcinomas, the ratiobetween expression of E-cadherin and collagenase type IV wasa highly significant predictor of survival. Importantly, in eachdisease, the ratio cutoff between groups differed, as would beexpected from tissues with a different biological phenotype. Inlung cancer, we found the ratio of 2.0 to separate between twodistinct prognostic groups. Specifically, the MMP:E-cadherinratio of 2 corresponded to the third quartile of the ratio. Signif-icant results (atP , 0.01) of the predictive value of the MMP:E-cadherin ratio were consistently found for overall survivalusing all cutpoints$1.8 and for disease-free survival using allcutpoints$1.7. The correctedPs remained statistically signifi-cant after applying the method proposed by Altmanet al.(42) tocorrect for the choice of optimal cutpoint usinge of 5 and 10%.

The observed results are consistent with the biologicalroles that E-cadherin and collagenase type IV play in the met-astatic cascade. E-cadherin is a cell surface glycoprotein in-volved in calcium-dependent homotypic cell-to-cell cohesion(43). It is localized at the epithelial junction complex and isresponsible for the organization, maintenance, and morphogen-esis of epithelial tissues. Reduced levels of E-cadherin areassociated with a decrease in cellular/tissue differentiation andincreased histological grade in different epithelial neoplasms(44 – 46). Transfection of E-cadherin-encoding cDNA intoinvasive cancer cells has been shown to inhibit their inva-siveness (47).

Once cells detach from the primary tumor, they mustinvade the host stroma if they are to metastasize (48, 49).Degradation of blood vessel basement components, especiallytype IV collagen, is one of the necessary steps in metastasis. Thelevels ofMr 72,000 andMr 92,000 type IV collagenase in humanand rodent neoplasms directly correlate with invasion and me-tastasis, and specific inhibitors of MMPs have been shown toinhibit tumor cell invasion (48–52). Thus, a decrease in theexpression of E-cadherin and increase in collagenase type IVactivity would enhance detachment of tumor cells and invasion.

In this limited series, the disease-free survival of patientswith T1 did not differ from those with T2 disease. Although it isencouraging that the MMP:E-cadherin ratio also did not differbetween these groups, only a larger series can determine the truesignificance of the analysis.

It is interesting that in our study, the expression of theproangiogenic molecules (bFGF, VEGF/VPF, and IL-8) did notcorrelate with survival (12, 52, 53). The regulation of thesegenes might be at the posttranscriptional level, necessitating

Fig. 3 Kaplan-Meier survival curvesversusMMP:E-cadherin expres-sion ratio in resected stage I NSCLCs. The overall survival differedsignificantly between tumors with a MMP:E-cadherin ratio of,2.0 andthose with a ratio higher than or equal to 2.0 (A;P 5 0.0002). Signif-icant differences (P5 0.0001) in disease-free survival are shown inB.




analysis by immunohistochemistry, which we did not do. Alter-natively, their expression could be more critical in more ad-vanced disease presentations.

The validation of these data requires a prospective study oranalysis of an independent set of NSCLC patients. Other poten-tial markers of prognosis in this population, including thoseinvolved in neoplastic angiogenesis, should also be useful inidentifying individuals most likely to benefit from adjuvanttherapies. At present, there are several MMP inhibitors availablein clinical practice, and one could conceive of a clinical trialwhere these agents are given to patients manifesting a highMMP:E-cadherin ratio at the time of their resection.

In summary, we used the ISH technique to examine theconcurrent expression of metastasis-related genes in formalin-fixed, paraffin-embedded specimens of resected stage I lungcarcinomas from patients undergoing curative surgery. Usingquantification of gene expression by colorimetric scanning andstandardization to nonpathological lung tissue, we conclude thatthe ratio of MMP-2 and MMP-9 expression to E-cadherin ex-pression at the periphery of lung cancers can accurately andsignificantly predict outcome (P 5 0.00003) in individual pa-tients with early stage disease.

AcknowledgmentsWe thank Walter Pagel for critical review and comments and Lola

Lopez and Bich Tran for expert assistance in the preparation of themanuscript.

References1. Martini, N., McCaughan, B. C., and McCormack, P. Lobectomy forstage I lung cancer.In: C. F. Kittle (ed.), Current Controversies inThoracic Surgery, pp. 171–174. Philadelphia: W. B. Saunders Com-pany, 1986.2. Williams, D. E., Pairolero, P. C., Davis, C. S., Bernatz, P. E., Payne,W. S., Taylor, W. F., Uhlenhopp, M. A., and Fontana, R. S. Survival ofpatients surgically treated for stage I lung cancer. J. Thorac. Cardiovasc.Surg.,82: 70–76, 1981.3. Teramachi, M., Miyamoto, N., Sasaka, T., Nakamura, T., and Kita-mura, F. Results of surgery in non-small cell lung cancer. Bull. ChestDis. Res. Inst. Kyoto Univ.,23: 48–53, 1990.4. Naruke, T., Goya, T., Tsuchiya, R., and Suemasu, K. Prognosis andsurvival in resected lung carcinoma based on the new internationalstaging system. J. Thorac. Cardiovasc. Surg.,96: 440–447, 1988.5. Slebos, R. J., Kibbelaar, R. E., Dalesio, O., Kooistra, A., Stam, J.,Meijer, C. J., Wagenaar, S. S., Vanderschueren, R. G., van Zandwijk,N., and Moot, W. J. K-rasoncogene activation as a prognostic markerin adenocarcinoma of the lung. N. Engl. J. Med.,323: 561–565, 1990.6. Lee, J. S., Ro, J. Y., Sahin, A. A., Hong, W. K., Brown, B. W.,Mountain, C. F., and Hittelman, W. N. Expression of blood-groupantigen A–a favorable prognostic factor in non-small-cell lung cancer.N. Engl. J. Med.,324: 1084–1090, 1991.7. Pezzella, F., Turley, H., Kuzu, I., Tungekar, M. F., Dunnill, M. S.,Pierce, C. B., Harris, A., Gatter, K. C., and Mason, D. Y. Bcl-2 proteinin non-small-cell lung carcinoma. N. Engl. J. Med.,329: 690–694,1993.8. Harpole, D. H., Herndon, J. E., and Young, W. G. Stage I non-smallcell lung cancer: a multivariate analysis of treatment methods andpatterns of recurrence. Cancer (Phila.),76: 787–796, 1995.9. Kwiatkowski, D. J., Harpole, D. H., Jr., Godleski, J., Herndon, J. E.,II, Shieh, D. B., Richards, W., Blanco, R., Xu, H. J., Strauss, G. M., andSugarbaker, D. J. Molecular pathologic substaging in 244 stage I non-small-cell lung cancer patients: clinical implications. J. Clin. Oncol.,16:2468–2477, 1998.

10. Mehdi, S. A., Tatum, A. H., Newman, N. B., Gamble, G. P., Etzell,J. E., Weidner, H., Kern, J. A., Sorscher, S. M., Kohman, L. J., andGraziano, S. L. Prognostic markers in resected stage I and II non-smallcell lung cancer: an analysis of 260 patients with 5 year follow-up. Clin.Lung Cancer,1: 59–67, 1999.

11. Pastorino, U., Andreola, S., Tagliabue, E., Pezzella, F., Incarbone,M., Sozzi, G., Buyse, M., Menard, S., Pierotti, M., and Rilke, F.Immunocytochemical markers in stage I lung cancer: relevance toprognosis. J. Clin. Oncol.,15: 2858–2865, 1997.

12. Fontanini, G., Boldrini, L., Chine, S., Pisaturo, F., Basolo, F.,Calcinai, A., Lucchi, M., Mussi, A., Angeletti, C. A., and Bevilacqua, G.Expression of vascular endothelial growth factor mRNA in non-small-cell lung carcinomas. Br. J. Cancer,79: 363–369, 1999.

13. Fontanini, G., Lucchi, M., Vignati, S., Mussi, A., Ciardiello, F., DeLaurentis, M., De Placido, S., Basolo, F., and Angeletti, C. A. Angio-genesis as a prognostic indicator of survival in non-small-cell lungcarcinoma: a prospective study. J. Natl. Cancer Inst.,89: 881–886,1997.

14. Fontanini, G., Bigini, D., Vignati, S., Basolo, F., Mussi, A., Lucchi,M., Chine, S., Angeletti, C. A., Harris, A. L., and Bevilacqua, G.Microvessel count predicts metastatic disease and survival in non-smallcell lung cancer. J. Pathol.,177: 57–63, 1995.

15. Fidler, I. J. Critical factors in the biology of human cancer metas-tasis: twenty-eighth G. H. A. Clowes Memorial Award Lecture. CancerRes.,50: 6130–6138, 1990.

16. Fidler, I. J., Kumar, R., Bielenberg, D. R., and Ellis, L. M. Molec-ular determinants of angiogenesis in cancer metastasis. Cancer J. Sci.Am., 4: S58–S66, 1998.

17. Fidler, I. J. Critical determinants of cancer metastasis: rationale fortherapy. Cancer Chemother. Pharmacol.,43: S3–S10, 1999.

18. Radinsky, R., Bucana, C. D., Ellis, L. M., Sanchez, R., Cleary,K. R., Brigati, D. J., and Fidler, I. J. A rapid colorimetricin situmessenger RNA hybridization technique for analysis of epidermalgrowth factor receptor in paraffin-embedded surgical specimens ofhuman colon carcinomas. Cancer Res.,53: 937–943, 1993.

19. Fidler, I. J., and Radinsky, R. Genetic control of cancer metastasis.J. Natl. Cancer Inst.,82: 166–168, 1990.

20. Bucana, C. D., Fabra, A., Sanchez, R., and Fidler, I. J. Differentpatterns of macrophage infiltration into allogeneic-murine and xenoge-neic-human neoplasms growing in nude mice. Am. J. Pathol.,141:1225–1236, 1992.

21. Kitadai, Y., Ellis, L. M., Tucker, S. L., Greene, G. F., Bucana,C. D., Cleary, K. R., Takahashi, Y., Tahara, E., and Fidler, I. J.Multiparametricin situ mRNA hybridization analysis to predict diseaserecurrence in patients with colon carcinoma. Am. J. Pathol.,149:1541–1551, 1996.

22. Kitadai, Y., Ellis, L. M., Takahashi, Y., Bucana, C. D., Anzai, H.,Tahara, T., and Fidler, I. J. Multiparametricin situ messenger RNAhybridization analysis to detect metastasis-related genes in surgicalspecimens of human colon carcinomas. Clin. Cancer Res.,1: 1095–1102, 1995.23. Kitadai, Y., Bucana, C. D., Ellis, L. M., Anzai, H., Tahara, E., andFidler, I. J. In situ mRNA hybridization technique for analysis ofmetastasis-related genes in human colon carcinoma cells. Am. J. Pathol.,147: 1238–1247, 1995.24. Anzai, H., Kitadai, Y., Bucana, C. D., Sanchez, R., Omoto, R., andFidler, I. J. Expression of metastasis-related genes in surgical specimensof human gastric cancer can predict disease recurrence. Eur. J. Cancer,34: 558–565, 1998.25. Greene, G. F., Kitadai, Y., Pettaway, C. A., von Eschenbach, A. C.,and Fidler, I. J. Correlation of metastasis-related gene expression withmetastatic potential in human prostate carcinoma cells implanted innude mice using anin situ messenger RNA hybridization technique.Am. J. Pathol.,150: 1571–1582, 1997.26. Kuniyasu, H., Ellis, L. M., Evans, D. B., Abbruzzese, J. L., Feno-glio, C. J., Bucana, C. D., Cleary, K. R., Tahara, E., and Fidler, I. J.Relative expression of E-cadherin and type IV collagenase genes pre-




dicts disease outcome in patients with resectable pancreatic carcinoma.Clin. Cancer Res.,5: 25–33, 1999.27. Collier, I. E., Wilhelm, S. M., Eisen, A. Z., Marmer, B. L., Grant,G. A., Seltzer, J. L., Kronberger, A., He, C. S., Bauer, E. A., andGoldbert, G. I. H-rasoncogene-transformed human bronchial epithelialcells (TBE-1) secrete a single metalloprotease capable of degradingbasement membrane collagen. J. Biol. Chem.,263: 6579–6587, 1988.28. Singh, R. K., Gutman, M., Radinsky, R., Bucana, C. D., and Fidler,I. J. Expression of interleukin 8 correlates with the metastatic potentialof human melanoma cells in nude mice. Cancer Res.,54: 3242–3247,1994.29. Singh, R. K., and Fidler, I. J. Regulation of tumor angiogenesis byorgan-specific cytokines. Curr. Top. Microbiol. Immunol.,213: 1–11,1996.30. Li, L. M., Nicolson, G. L., and Fidler, I. J. Directin vitro lysis ofmetastatic tumor cells by cytokine-activated murine vascular endothelialcells. Cancer Res.,51: 245–254, 1991.31. Tischer, E., Mitchell, R., Hartman, T., Silva, M., Gospodarowicz,D., Fiddes, J. C., and Abraham, J. A. The human gene for vascularendothelial growth factor. Multiple protein forms are encoded throughalternative exon splicing. J. Biol. Chem.,266: 11947–11954, 1991.32. Matsuyama, K., Chiba, Y., Sasaki, M., Tahaka, H., Muraoka, R.,and Tanigawa, N. Tumor angiogenesis as a prognostic marker in oper-able non-small cell lung cancer. Ann. Thorac. Surg.,65: 1405–1409,1998.33. Bussemakers, M. J., van Bokhoven, A., Mees, S. G., Kemler, R.,and Schalken, J. A. Molecular cloning and characterization of thehuman E-cadherin cDNA. Mol. Biol. Rep.,17: 123–128, 1993.34. Pearson, W. R., and Lipman, D. J. Improved tools for biologicalsequence comparison. Proc. Natl. Acad. Sci. USA,85: 2444–2448,1988.35. Park, C-S., Manahan, I. J., and Brigati, D. J. Automated molecularpathology: one hourin situ DNA hybridization. J. Histotechnol.,14:219–229, 1991.36. McCurley, C. R., Shivers, R. R., and Del Maestro, R. F. Quantita-tive comparison of the morphology of the microvasculature of primarylung lesions and metastatic brain tumours. J. Submicrosc. Cytol. Pathol.,30: 257–269, 1998.37. Caruthers, M. H., Beaucage, S. L., Efcavitch, J. W., Fisher, E. F.,Goldman, R. A., de Haseth, P. L., Mandecki, W., Matteucci, M. D.,Rosendahl, M. S., and Stabinsky, Y. Chemical synthesis and biologicalstudies on mutated gene-control regions. Cold Spring Harbor Symp.Quant. Biol.,47: 411–418, 1983.38. Kaplan, E. L., and Meier, P. Nonparametric estimation from incom-plete observation. J. Am. Stat. Assoc.,53: 457–481, 1958.39. Bucana, C. D., Radinsky, R., Dong, Z., Sanchez, R., Brigati, D. J.,and Fidler, I. J. A rapid colorimetricin situ mRNA hybridizationtechnique using hyperbiotinylated oligonucleotide probes for analysis ofmdr1 in mouse colon carcinoma cells. J. Histochem. Cytochem.,41:499–506, 1993.40. Kitadai, Y., Haruma, K., Sumii, K., Yamamoto, S., Ue, T.,Yokozaki, H., Yasui, W., Ohmoto, Y., Fidler, I. J., Tahara, E., andKajiyama, G. Expression of interleukin-8 correlates with vascularity inhuman gastric carcinomas. Am. J. Pathol.,152: 93–100, 1998.41. Fidler, I. J., and Kripke, M. L. Metastasis results from preexistingvariant cells within a malignant tumor. Science (Washington DC),197:893–895, 1977.

42. Altman, D. G., Lausen, B., Sauerbrei, W., and Schumacher, M.Dangers of using “optimal” cutpoints in the evaluation of prognosticfactors. J. Natl. Cancer Inst.,86: 829–835, 1994.

43. Takeichi, M. Cadherin cell adhesion receptors as a morphogeneticregulator. Science (Washington DC),251: 1451–1455, 1991.

44. Takeichi, M. The cadherins: cell-cell adhesion molecules control-ling animal morphogenesis. Development (Camb.),102: 639–655,1988.

45. Shimoyama, Y., Hirohashi, S., Hirano, S., Noguchi, M., Shimosato,Y., Takeichi, M., and Abe, O. Cadherin cell-adhesion molecules inhuman epithelial tissues and carcinomas. Cancer Res.,49: 2128–2133,1989.

46. Shimoyama, Y., and Hirohashi, S. Expression of E- and P-cadherinin gastric carcinomas. Cancer Res.,51: 2185–2192, 1991.

47. Vleminckx, K., Vakaet, L., Jr., Mareel, M., Fiers, W., and van Roy,F. Genetic manipulation of E-cadherin expression by epithelial tumorcells reveals an invasion suppressor role. Cell,66: 107–119, 1991.

48. Liotta, L. A., Tryggvason, K., Garbisa, S., Hart, I., Foltz, C. M., andShafie, S. Metastatic potential correlates with enzymatic degradation ofbasement membrane collagen. Nature (Lond.),284: 67–68, 1980.

49. Liotta, L. A., and Stetler-Stevenson, W. G. Metalloproteinases andcancer invasion. Semin. Cancer Biol.,1: 99–106, 1990.

50. Matrisian, L. M., and Bowden, G. T. Stromelysin/transin and tumorprogression. Semin. Cancer Biol.,1: 107–115, 1990.

51. Schipper, J. H., Frixen, U. H., Behrens, J., Unger, A., Jahnke, K.,and Birchmeier, W. E-cadherin expression in squamous cell carcinomasof head and neck: inverse correlation with tumor dedifferentiation andlymph node metastasis. Cancer Res.,51: 6328–6337, 1991.

52. Oka, H., Shiozaki, H., Kobayashi, K., Inoue, M., Tahara, H., Koba-yashi, T., Takatsuka, Y., Matsuyoshi, N., and Hirano, S. Expression ofE-cadherin cell adhesion molecules in human breast cancer tissues andits relationship to metastasis. Cancer Res.,53: 1696–1701, 1993.53. Imoto, H., Osaki, T., Taga, S., Ohgami, A., Ichiyoshi, Y., andYasumoto, K. Vascular endothelial growth factor expression in non-small-cell lung cancer: prognostic significance in squamous cell carci-noma. J. Thorac. Cardiovasc. Surg.,115: 1007–1014, 1998.54. Takanami, I., Tanaka, F., Hashizume, T., and Kodaira, S. Vascularendothelial growth factor and its receptor correlate with angiogenesisand survival in pulmonary adenocarcinoma. Anticancer Res.,17: 2811–2814, 1997.55. Singh, R. K., Bucana, C. D., Gutman, M., Fan, D., Wilson, M. R.,and Fidler, I. J. Organ site-dependent expression of basic fibroblastgrowth factor in human renal cell carcinoma cells. Am. J. Pathol.,145:365–374, 1994.56. Rogelj, S., Weinberg, R. A., Fanning, P., and Klagsbrun, M. Basicfibroblast growth factor fused to a signal peptide transforms cells.Nature (Lond.),331: 173–175, 1988.57. Leung, D. W., Cachianes, G., Kuang, W. J., Goeddel, D. V., andFerrera, N. Vascular endothelial growth factor is a secreted angiogenicmitogen. Science (Washington DC),246: 1306–1309, 1989.58. Matsushima, K., Morishita, K., Yoshimura, T., Lavu, S., Koba-yashi, Y., Lew, W., Appella, E., King, H. F., Leonard, E. J., andOppenheim, J. J. Molecular cloning of a human monocyte-derivedneutrophil chemotactic factor (MDNCF) and the induction of MDNCFmRNA by interleukin 1 and tumor necrosis factor. J. Exp. Med.,167:1883–1893, 1988.




2000;6:790-797. Clin Cancer Res Roy S. Herbst, Seiji Yano, Hiroki Kuniyasu, et al. Cell Lung CarcinomaGenes Predicts Outcome in Patients with Stage I Non-Small Differential Expression of E-cadherin and Type IV Collagenase

Updated version

http://clincancerres.aacrjournals.org/content/6/3/790

Access the most recent version of this article at:

Cited articles

http://clincancerres.aacrjournals.org/content/6/3/790.full#ref-list-1

This article cites 52 articles, 20 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/6/3/790.full#related-urls

This article has been cited by 19 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://clincancerres.aacrjournals.org/content/6/3/790To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/content/6/3/790.full#ref-list-1

http://clincancerres.aacrjournals.org/content/6/3/790.full#related-urls

http://clincancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



differential expression of e-cadherin and type iv...

Documents